determination of cyclamate in foods

FOOD COMPOSITION AND ADDITIVES

Determination of Cyclamate in Foods by UltraperformanceLiquid Chromatography/Tandem Mass Spectrometry

ROBERT SHERIDAN and THOMAS KING

New York State Department of Agriculture and Markets, Food Laboratory, Building 7, State Office Campus, Albany, NY

12235

A highly sensitive and selective method that

requires minimal sample preparation was

developed for the confirmation and quantitation of

cyclamate in a variety of foods by high-performance

liquid chromatography/tandem mass spectrometry

(HPLC/MS/MS). Sample preparation consisted of

homogenization followed by extraction and

dilution of cyclamate with water. HPLC separation

was achieved using a bridged ethyl hybrid C18

high-pressure column with a mobile phase

consisting of 0.15% acetic acid and methanol.

Under electrospray ionization negative conditions,

quantitation was achieved by monitoring the

fragment m/z = 79.7 while also collecting parent

ion m/z = 177.9. Two food matrixes, diet soda and

jelly, were subjected to a validation procedure in

order to evaluate the applicability of the method.

The cyclamate limit of detection for both matrixes

was determined to be 0.050 �g/g with a limit of

quantitation of 0.150 �g/g. The correlation

coefficient of the calibration curves was �0.9998

from 0.0005 to 0.100 �g/mL. The method has been

used for the determination of cyclamate in several

foods and the results are presented.

Sugar substitutes are widely used in the production of

various types of foods. The need for noncaloric

sweeteners for diabetics as well as for individuals

concerned with carbohydrate consumption means that there

will be a great demand for such products. Cyclamate

(N-cyclohexylsulfamate) is a noncaloric sweetener used in

many foods in many countries. It is 30 times sweeter than

sucrose and its sweetening effectiveness increases when it is

used in combination with other artificial sweeteners such as

saccharin (1, 2).

Although cyclamate is allowed for use in food in many

countries, it is not approved for use in food in the United

States because of concerns that it and its metabolite may cause

bladder cancer. Subsequent studies have failed to corroborate

the link to bladder cancer; however, the U.S. cyclamate ban

remains. In recent years an increasing amount of food sold in

the United States is imported from regions of the world where

cyclamate is permitted. In those countries, there are no

regulations regarding the amount of cyclamate allowed in

food, and cyclamate has been detected in foods that typically

require added sweeteners at concentrations as high as

5.08 g/kg. For this reason, a simple, selective, and sensitive

analytical method is needed for detecting cyclamate in foods

over a large range of concentrations.

One of the earliest techniques for cyclamate determination

in foods is high-performance liquid chromatographic (HPLC)

separation followed by UV or conductivity detection.

However, because cyclamate lacks a chromophore necessary

for UV detection, a complex and time-consuming

derivatization procedure is required (3–5). This approach

often involves hazardous reagents or specialized hardware

such as post-column extractors. An alternative detection

scheme for cyclamate detection is the use of an

electrochemical detector in conductivity mode (6). Several

gas chromatographic methods have been reported, but they

too require derivatization of cyclamate prior to analysis (7, 8).

Capillary electrophoresis provides an option where separation

and detection can be performed without derivatization;

however extraction and cleanup steps may be necessary in

order to remove potential interferences (9, 10). Perhaps the

most common technique for cyclamate analysis currently is

flow injection followed by spectrophotometric or

electrochemical detection (11–15). These methods typically

incorporate an in-line derivatization procedure necessary for

analyte detection. When the derivitization method is based

upon the measurement of a reducing agent using the Griess

reaction, the presence of ascorbic acid can interfere with

cyclamate detection (12, 13). Atomic absorption detection can

also be used if the cyclamate is first oxidized and precipitated

with lead nitrate (16). One technique of detecting cyclamate

without derivatization involves the addition of a

chromatographic dye (methyl-red) to the mobile phase

followed by UV-Vis absorbance detection (17). Few

published methods, however, use mass spectrometry (MS) as

a detection scheme. Single-stage MS can provide analyte

structural information, greatly improving method selectivity

over most other strategies (2, 18); however, tandem mass

spectrometry (MS/MS) would provide added selectivity and

sensitivity over single-stage MS.

SHERIDAN & KING: JOURNAL OF AOAC INTERNATIONAL VOL. 91, NO. 5, 2008 1095

Received February 12, 2008. Accepted by SG April 30, 2008.Corresponding author’s e-mail: [email protected]

To our knowledge, no method has been published which

uses LC/MS/MS for the determination of cyclamate in food.

This technique would provide improved selectivity and

confirmation confidence over other methods. We present a

method which uses a minimum of sample preparation

followed by HPLC/MS/MS analysis of cyclamate in several

foods. The selectivity of MS/MS greatly reduces the

possibility of analyte interference with co-extractives, and due

to the inherent sensitivity of this technique, the extracted

sample can be diluted while still providing low detection

limits. Results of an ongoing cyclamate monitoring survey, as

well as validation data, are presented here.

Experimental

Apparatus

(a) LC system.—Acquity HPLC, Waters Acquity binary

solvent manager with sample manager, column manager, and

in-line degasser (Waters, Milford, MA).

(b) Mass spectrometer.—Quattro Premier XE (Waters).

(c) Controlling software.—MassLynx 4.1 (Waters).

(d) LC column and guard column.—Bridged ethyl hybrid

(BEH) C18, 1.7 �m, 1.0 � 100 mm (Waters).

(e) Centrifuge.—Thermo Electron (Milford, MA).

Materials and Reagents

(a) Cyclamate standard.—Sigma (St. Louis, MO). A

1 mg/mL cyclamate stock standard was prepared by

dissolving 0.010 g neat cyclamate standard in 10 mL 50%

deionized water–methanol. All working standards are made in

deionized water without matrix matching. A standard curve

1096 SHERIDAN & KING: JOURNAL OF AOAC INTERNATIONAL VOL. 91, NO. 5, 2008

Table 1. MS conditions for 2 cyclamate transition ions

RT,min

Iontransition, m/z

Conevoltage

Dwelltime, s

Collisionenergy, eV

4.84 Quantitative

ion

177.9 > 79.7 45.00 0.10 25.0

Qualitative

ion

177.9 > 177.9 45.00 0.10 12.0

Figure 1. Cyclamate standard, 10 ng/mL, with both transitions (177.9 > 79.7 and 177.9 > 177.9).

consisting of at least 5 points was generated to quantify each

group of samples. Typically, standards of 0.001, 0.005, 0.010,

0.050, and 0.100 �g/mL (ppm) were used, and all samples

found to contain cyclamate were diluted into this range for

quantitation.

(b) Disposable centrifuge tubes.—50 mL (Fisher

Scientific, Worcester, MA).

(c) Disposable syringe filters.—0.2 �m nylon, 25 mm

diameter (Whatman, Florham Park, NJ).

(d) Deionized water source.—Barnstead Nanopure II

(Dubuque, IA).

(e) Disposable syringes.—3 cc Luer Lock (Kendall,

Mansfield, MA).

(f) Polyethylene bags.—Clear View Bag Co. (Albany, NY).

(g) Methanol.—J.T. Baker (Phillipsburg, NJ); HPLC

RESI-analyzed.

(h) HPLC mobile phase.—A: 0.15% acetic acid;

B: 0.15% acetic acid in methanol.

Instrument Conditions

(a) Column temperature.—30�C.

(b) LC program.—The flow was held at 0.080 mL/min

throughout the 11 min run. The initial mobile phase conditions

were 90% A and held for 1 min; by 6 min the mobile phase

consisted of 2.0% A and was held at this condition until

7.0 min. From 7.0 to 7.5 min, the mobile phase was ramped to

the original conditions where they were held for the remainder

of the 11 min program.

(c) Injection volume.—10 �L.

(d) Sample temperature.—10�C.

(e) Divert valve program.—On from 0 to 1 min and from 8

to 11 min.

(f) MS source temperature.—100�C.

(g) Desolvation temperature.—250�C.

(h) Cone gas flow.—300 L/h.

(i) Desolvation gas flow.—900 L/h.

(j) MS mode.—Multireaction monitoring negative.

(k) Injection mode.—Full loop.

A description of MS acquisition conditions is presented in

Table 1.

Sample Preparation/Extraction

Solid food samples were homogenized in a food processor.

Food types such as jelled fruit and gummy materials were

placed in a plastic bag in a –80�C freezer overnight. The

samples were then removed from the freezer, immediately

placed in a polyethylene bag, and crushed to a particle size as

small as possible. A 1.0 g liquid or well-homogenized food

sample was weighed into a 50 mL disposable centrifuge tube,

and 10 mL distilled deionized water was added. The sample

was mixed on a Vortex mixer for 10 min. It was then

centrifuged for 10 min, and 3 mL of the supernatant was

filtered through a 0.2 �m disposable syringe filter. A 1.0 mL

volume of the filtered supernatant was then added to a 10 mL

volumetric flask and diluted to volume with deionized water.

This final solution was transferred to an autosampler vial for

LC/MS/MS analysis.


Figure 2. Daughter ions produced from the parent m/z = 177.9 cyclamate.

Results and Discussion

Chromatography

Initially, LC/MS cyclamate analysis was performed using a

standard HPLC system coupled to a single-quadrupole MS.

The column used was an Xterra MS C18, 3.5 �m analytical

column with several flow rates applied. This method

produced a poorly resolved cyclamate peak with large peak

width and an inconsistent retention time. Several sample

matrixes also produced interference for m/z = 178 and 79.7

used to identify cyclamate. For these reasons, an alternative

method was sought.

An improvement in chromatographic resolution,

sensitivity, and selectivity was achieved with the use of

ultraperformance LC/MS/MS. This system consisted of a

BEH C18 column with a particle size of 1.7 �m installed on a

Waters Acquity HPLC with a mobile phase of A = 0.15%

acetic acid and B = 0.15% acetic acid in methanol. With a flow

rate of 0.08 mL/min, the mobile phase gradient began at 90%

aqueous and progressed to 98% organic at 7 min. Under these

conditions, cyclamate was well resolved at 4.8 min and the

retention time was stable. The improvement in selectivity and

sensitivity was due to the use of MS/MS; however, it is not

clear why the higher-pressure LC system gave much better

chromatographic resolution. It is possible that the smaller

particle size, narrower column dimensions, and higher

operating pressure combined to improve cyclamate peak

resolution. Figure 1 shows typical chromatograms of the 2

cyclamate transitions generated from a 0.010 �g/mL

cyclamate standard.

Sample Preparation

Because cyclamate is water-soluble and the analysis

technique is extremely selective, a minimum of preparation is

necessary. The aqueous extract can be easily centrifuged,

filtered, and diluted 200-fold, at which point the co-extracted

matrix does not interfere with the analysis. No difficulties with

emulsions being formed during sample extraction were

observed; however, the method was not tested with high-fat

food. The sensitivity of MS/MS allows for low detection

limits even after a large dilution, and no matrix-induced signal

suppression is evident. For this reason, it is not necessary to

matrix-match standards. No sample preparation difficulties

were encountered using this method with foods as varied as

soda, plum butter, dried fruit, candy, and cake.

Mass Spectrometry

All MS conditions were optimized by a flow injection of

cyclamate solution into the MS source. HPLC mobile phase

was pumped into the MS source through a Tee connector

during the flow of cyclamate solution in order to mimic

analytical run conditions. For this experiment, the mobile

phase was set to the initial conditions of the HPLC program,

which was 10% methanol at a rate of 0.080 mL/min. A10 ppm

cyclamate standard in deionized water was injected into this

flow at 20 �L/min. The MS source conditions were set as

described above and the collision gas flow was turned off. The

MS scan type was set to MS1 in order to collect full-scan data.

In this mode, the collision cell and MS2 are used to pass ions

to the detector. Being an anion, the strongest signal for

cyclamate was seen in negative mode, and with a cone voltage

of 45, the quasimolecular ion m/z = 177.9 was observed. With

the collision gas on and the instrument in daughter-scan mode,

m/z = 177.9 was isolated and fragmented while the collision

energy was varied. As the collision energy was increased, a


Figure 3. Proposed product ion formation schemefrom cyclamate via electrospray ionization/tandemmass spectrometry.

Table 2. Recovery data from limit of quantitation (LOQ)

method spikes

Spike level

Recovery, %

Diet soda JellyDried

tamarindStrawberries

in syrup

LOQ 1 70.3 81.5 78.2 74.7

LOQ 2 70.2 76.9 73.4 77.4

LOQ 3 72.0 85.4 76.5 74.2

10 � LOQ 1 89.2 75.5 81.0 81.5

10 � LOQ 2 88.6 75.2 86.4 81.1

10 � LOQ 3 90.8 78.6 78.2 83.6

100 � LOQ 1 98.9 84.1 86.4 89.8

100 � LOQ 2 94.9 83.2 85.1 79.0

100 � LOQ 3 94.5 85.4 84.3 77.9

1000 � LOQ 1 99.7 86.1 92.6 93.9

1000 � LOQ 2 100 84.5 92.3 93.7

1000 � LOQ 3 96.1 89.7 91.9 94.0

Mean 88.77 82.18 83.86 83.40

SD 11.46 4.64 6.49 7.54

CV, % 12.9 5.7 7.7 9.0

strong signal at m/z = 79.7 was observed and a further increase

in collision energy resulted in a reduction of m/z = 79.7 signal

strength with no other ions produced (Figure 2). A further

increase in collision energy only resulted in a reduction in

signal abundance without producing novel fragments. The

79.7 ion was the same fragment observed by others using

single-stage MS where it was produced in the ion source (2).

Single-stage MS, however, is susceptible to matrix

interference, whereas MS/MS is much more selective owing

to the isolation of the parent ion prior to fragmentation.

Because only one parent-daughter transition was

generated, we chose to also collect the unfragmented

m/z = 177.9. Ideally, at least 2 daughter fragments should be

produced from a given parent ion in order to satisfy generally

accepted requirements of LC/MS/MS confirmation (19);

however, in some cases this is not possible. While not a

traditional ion ratio measurement, the abundance ratio of

m/z = 177.9 > 79.7 to 177.9 > 177.9 can be used to improve

confirmation confidence over simply observing the presence

of m/z = 177.9 > 79.7. This situation is sometimes unavoidable

when dealing with small molecules. We chose to set

acceptance limits of 10% when comparing quantitation ion

(79.7) to qualifier ion (177.9) for confirming the presence of

cyclamate. Figure 3 represents the proposed fragmentation

scheme for generation of the 79.7 ion. Along with the ion ratio

criteria, the retention time of the analyte in the sample must

match that in the standard within 5% in order for the sample to

be considered confirmed.

One technique for maintaining consistent instrument

response over the course of many injections is through the use

of the flow divert valve. By sending the LC flow to waste for

the initial 1 min, much of the water-soluble co-extracted

potential interference is kept out of the MS source. This

increases the number of sample injections that can be made

before the MS source cone needs to be cleaned, thereby

increasing the number of samples that can be analyzed at

one time.

Method Validation

For the purposes of method validation, the limit of

detection (LOD) was set at 50 ng/g (ppb) for all food types.


Table 3. Results from pomegranate soda with incurred

cyclamate, extracted and analyzed 10 times

Method repeatability

Run Concn, �g/g

1 198

2 205

3 202

4 221

5 204

6 205

7 202

8 207

9 212

10 223

Mean 208

SD 8.3

CV, % 3.98

Table 4. Commodities found to contain cyclamate

Commodity type No. detected Country of origin

High concentration

found, �g/g

Low concentration

found, �g/g

Plum butter 1 Hungary <0.15 0.103

Pitted cherries 1 Bulgaria 0.2 0.2

Dried prunes 4 China, Taiwan 42499 37500

Mangos in syrup 2 China, Taiwan 439 378

Preserved peach 2 China 3939 2.63

Preserved olive 6 China, Hong Kong 3710 25.9

Preserved plum 15 China, Hong Kong, Taiwan 41253 0.56

Pomegranate soda 1 Multiple countries 241 241

Sweetened kumquat 1 China <0.15 <0.15

Sweet orange 1 China 15300 15300

Grape tomato 1 China 1850 1850

Dried fruit candy 3 China 56300 35500

Dried beans 1 China 664 664

Strawberry cake 1 Taiwan 0.34 0.34


Figure 4. Transitions for (a) yam jelly containing no cyclamate; (b) 50 ppb cyclamate standard; (c) preserved plumcontaining 1781 ppm cyclamate; and (d) dried plum containing 4390 ppm cyclamate.

This concentration produces a signal from both transitions

well above 10 times noise in all matrixes tested, and it is

unlikely that cyclamate would be added as a sweetener at

levels below this amount. Two sample types, jelly and diet

soda, were spiked at 0.050 �g/g in duplicate and extracted

according to the previously described method. Cyclamate was

confirmed in all LOD spikes with retention time and ion ratio

requirements met.

The limit of quantitation (LOQ) was determined to be

3 times the LOD, or 0.150 �g/g. At this concentration, both

ions produced signals well above 10 times noise and no

interferences were observed. To determine accuracy of

quantitation, 4 food matrixes (diet soda, jelly, dried tamarind,

and strawberries in syrup) were spiked at LOQ (0.15 ppm),

10 � LOQ (1.5 ppm), 100 � LOQ (15 ppm), and 1000 � LOQ

(150 ppm) in triplicate. All spikes were prepared by adding

0.15 mL cyclamate solution at the appropriate level. To obtain

the 1000 � LOQ at 150 ppm, 0.15 mL 1000 �g/mL cyclamate

was added to the 1.0 g blank sample. To obtain a higher

spiking level, cyclamate would have to be added in the pure

form, as 1000 �g/mL was the most concentrated solution

prepared. The spikes were extracted and analyzed in

accordance with the method presented. All recoveries were

between 70 and 100% and the overall average of these

recoveries was 84.6% (Table 2). The standard deviation (SD)

values for the diet soda, jelly, dried tamarind, and strawberries

in syrup were 11.46, 4.64, 6.49, and 7.54, respectively. The

coefficient of variation (CV) values for the diet soda, jelly,

dried tamarind, and strawberries in syrup were 12.9, 5.7, 7.7,

and 9.0%, respectively. The lower percent recovery observed

in some matrixes such as strawberries in syrup was most likely

due to higher ion suppression. Although some matrixes may

cause more suppression than others, leading to lower signal

and reduced recovery, recovery rates were still reasonable and

all confirmation criteria were met.

In order to evaluate the linearity of a typical cyclamate

standard curve, a 5 point curve was generated on 3 different

days, and the curve statistics were calculated. The correlation

coefficient for all curves was �0.9995, which included

standards from 0.001 to 0.1 �g/mL. All samples found to

contain cyclamate were diluted so that the quantitation

response was within the calibration curve range. Instrument

precision was evaluated by injecting one standard repeatedly

19 times, which resulted in a CV of 1.06%. Method

repeatability was demonstrated by extracting and analyzing a

pomegranate soda with incurred cyclamate 10 times. This

study resulted in an SD of 8.3 and a CV of 3.98% (Table 3).

Food Samples

In 2007, a surveillance program was begun in order to

determine the prevalence of cyclamate in foods sold in New

York state. Because cyclamate is not used in domestic food

production, the focus of this program was on imported

products that typically contain added sweeteners. In order to

increase the likelihood of identifying foods containing added

cyclamate, products were sampled for analysis based on the

suspicion that they contained undeclared sweeteners and that

these foods were imported from countries where cyclamate is

commonly used. In those countries, regulation may not

require the explicit labeling of all ingredients including those

that may not be permitted in other countries.

Using this method, cyclamate has been identified in a

number of imported products, including plum butter, dried

fruit, and pomegranate soda from China, Taiwan, Hungary,

and Bulgaria (Table 4). The concentration of cyclamate found

in samples varied greatly from a low of �0.15 �g/g in a

sweetened kumquat imported from China to a high of

56 300 �g/g in dried fruit candy imported from China. All of

the foods that contained �4% cyclamate were dried fruit or

dried fruit candy, which typically tastes very sweet.

Figure 4 illustrates the 2 cyclamate transitions of

m/z = 177.9 � 177.9 and 177.9 � 79.7 for (a) yam jelly

containing no cyclamate; (b) 0.050 �g/mL (ppm) standard of

cyclamate; (c) diluted extract of a preserved plum containing

1781 �g/g cyclamate; and (d) diluted extract of a dried plum

containing 4390 �g/g cyclamate. The ratio of transition area in

the 0.050 ppm standard was 37%, while that of the preserved

plum (c) and dried plum (d) were 38 and 37%, respectively.

Additional commodities that contained cyclamate included

sour cherries, dried prune, jarred peach, jarred olive, plum

butter, and pomegranate soda. None of these food types

presented problems with cyclamate detection even with the

minimal sample preparation used in this method. The wide

range of concentrations found illustrates the need for a large

linear dynamic range for quantitation. We were able to

repeatedly generate a curve from 0.001 to 0.1 �g/mL with a

correlation coefficient of 0.9999 or better, which means fewer

dilutions were necessary to quantify cyclamate.

Conclusions

Amethod for the detection and quantitation of cyclamate in

foods using HPLC/MS/MS has been developed. HPLC

provides good chromatographic resolution for both MS/MS

transitions monitored. The selectivity of MS/MS eliminates

the need for an elaborate sample cleanup procedure, and no

co-extracted interferences have presented problems with

cyclamate detection. This method has been validated and used

in a survey of imported foods suspected of containing

cyclamate. Cyclamate was detected in several food types,

including preserved fruit, soda, and juice. The brief sample

preparation combined with the short run time of the HPLC

method means that many samples can be completed in a day,

giving quantitation and instant confirmation.

References

(1) Calorie Control Council, http//www.caloriecontrol.

org/cyclamate.html

(2) Huang, Z., Jinyu, M., Chen, B., Zhang, Y., & Yoa, S. (2006)

Anal. Chem. Acta 555, 233–237

(3) Casals, I., Reixach, M., Amat, J., Fuentes, M., &

Serra-Marjem, L. (1996) J. Chromatogr. A 750, 397–402


(4) Lawrence, J.F., & Charbonneau, C.F. (1988) J. Assoc. Off.

Anal. Chem. 71, 934–937

(5) Zhu, Y., Guo, Y., Ye, M., & James, F.S. (2005) J.

Chromatogr. A 1085, 143–146

(6) Chen, Q., Mou, S., Liu, K., Yang, Z., & Ni, Z. (1997)

J. Chromatogr. A 771, 135–143

(7) Dalziel, J.A.W., Johnson, R.M., & Shenton, A.J. (1972)

Analyst 97, 719–722

(8) Conacher, H.B.S., & O’Brien, R.C. (1971) J. Assoc. Off.

Anal. Chem. 54, 1135–1137

(9) Horie, M., Ishikawa, F., Oishi, M., Shindo, T., Yasui, A., &

Ito, K. (2007) J. Chromatogr. A 1154, 423–428

(10) Thompson, C.O., Trenerry, V.C., & Kemmery, B. (1995) J.

Chromatogr. A 704, 203–210

(11) Cabero, C., Saurina, J., & Hernandez-Cassou, S. (1999) Anal.

Chim. Acta 381, 307–313

(12) Rocha, F.R.P., Rodenas-Torralba, E., Morales-Rubio, A., &

de la Guardia, M. (2005) Anal. Chim. Acta 547, 204–208

(13) Gouveia, S.T., Fatibello-Filho, O., & de Araujo Nobrega, J.

(1995) Analyst 120, 2009–2012

(14) Llamas, N.E., Di Nezio, M.S., Palomeque, M.E., &

Fernandez Band, B.S. (2005) Anal. Chim. Acta 539, 301–304

(15) Nikolelis, D.P., & Pantoulias, S. (2001) Anal. Chem. 73,

5945–5952

(16) Yebra, M.C., & Bermejo, P. (1998) Talanta 45, 1115–1122

(17) Choi, M.M.F., Hsu, M.Y., & Wong, S.L. (2000) Analyst 125,

217–220

(18) Koyama, M., Yoshida, K., Uchibori, N., Wada, I., Akiyama,

K., & Sasaki, T. (2005) J. Food Hyg. Soc. Jpn. 46, 72–78

(19) U.S. Food and Drug Administration (2003) Guidance for

Industry 118, Mass Spectrometry for Confirmation of the

Identity of Animal Drug Residues, Center for Veterinary

Medicine, Rockville, MD


determination of cyclamate in foods

Documents